MEMBRANE PROTEIN STRUCTURE AND DYNAMICS STUDIED BY

Bordignon, E., and Steinhoff, H.-J. (2007) Membrane protein structure and dynamics studied by site-directed spin labeling ESR. In: Hemminga, M.A.,and Berliner, L.J. (eds). ESR Spectroscopy in Membrane Biophysics.Springer Science and Business Media, New York, 129-164

CHAPTER 5

MEMBRANE PROTEIN STRUCTURE AND DYNAMICS STUDIED BY SITE-DIRECTED SPIN LABELING ESR

Enrica Bordignon and Heinz-Jürgen Steinhoff Universität Osnabrück, Osnabrück, Germany 1. INTRODUCTION ESR spectroscopy of site-directed spin-labeled biomolecules (Site-Directed Spin Labeling, SDSL) has emerged as a powerful method for studying the structure and conformational dynamics of proteins and nucleic acids under conditions relevant to function (for reviews see, e.g., Feix and Klug 1998; Hubbell et al. 1996; Hubbell et al. 1998, 2002). In this technique a spin-label side chain is introduced at a selected site via cysteine substitution mutagenesis followed by modification of the unique sulfhydryl group with a specific paramagnetic nitroxide reagent. The continuous wave (cw) ESR spectrum yields information about the nitroxide side chain mobility, solvent accessibility, the polarity of its immediate environment, and the distance between the nitroxide and another paramagnetic center in the protein. Hence, ESR data analysis of a series of spin-labeled variants of a given protein allows defining elements of secondary structure, including their solvent exposure, to characterize protein topography and to determine orientations of individual segments of the protein. A complete analysis allows modeling of protein structures with a spatial resolution at the level of the backbone fold (Hubbell et al. 1998, 2000; Koteiche and Mchaourab 1999; Mchaourab and Perozo 2000; Perozo et al. 1998; Wegener et al. 2001a). This method is applicable to any protein with a cloned gene that can be expressed. In particular, it has been shown to be very useful in studying large membrane proteins or protein complexes that are not amenable to NMR methods or do not crystallize. One of the most powerful properties of the method is its sensitivity to molecular dynamics: protein equilibrium fluctuations and conformational changes of functional relevance can be followed on a wide time scale ranging from picoseconds to seconds. The present chapter summarizes recent progress in the SDSL methodological approach, with special emphasis on application to the study of molecular dynamics, membrane protein structure, and protein–protein interaction.

2

ENRICA BORDIGNON AND HEINZ-JÜRGEN STEINHOFF

2. SPIN LABELING Elucidation of the structure and function of proteins by ESR spectroscopy has become increasingly important in recent years as technological advances have been made in the development of spectrometers and applicable methods. Such advances are also accompanied by novel chemistry approaches to design probes fulfilling desirable requirements in terms of incorporation into proteins, labeling efficiency, similarity with natural amino acid side chains, negligible perturbation of protein structure and function, and dynamical properties chosen ad hoc to monitor the backbone dynamics or to sample the microenvironment of the attached probe. There are two main approaches to modifying peptides or proteins with paramagnetic spin labels. A very well-established method consists in utilizing the sulfhydryl group of cysteine residues naturally present in the protein under investigation or engineered via site-directed mutagenesis to attach a variety of nitroxides to yield a spin-label side chain. This approach usually requires that the target protein possess only cysteine residues at the desired sites and that additional cysteines eventually present in the protein must be eliminated (cysteines are usually replaced by serines or alanines). Among the various spin labels utilized in literature works, the spin label (1-oxyl-2,2,5,5-tetramethylpyrroline-3-methyl)methanethiosulfonate (MTS) (Berliner et al. 1982) is most often used in SDSL studies due to its sulfhydryl specificity and its small molecular volume, similar to a tryptophane side chain (Figure 1A). The link between the piperidine-oxyl moiety and the protein backbone renders the label flexible, allowing native folding of the proteins when the spin label is attached (Figure 1D). Moreover, the unique dynamic properties of this spin-label side chain facilitate obtaining detailed structural information from the shape of its ESR spectrum (see §3.1). On the other hand, the structural variability characterizing the MTS spin label (the distance between the Cβ and the NO group varies between 4 and 8 Å depending on the binding site) does not unequivocally allow relating of the data obtained directly to the property of the native side chain. Additionally, in the case of distance determination between two labeled positions, only a distance distribution rather than a defined distance can be obtained, entailing the use of a higher number of spin-labeled sites and eventually of conformational searching approaches (Sale et al. 2005) for structural modeling of proteins based on SDSL ESR. The other approach includes incorporation of a paramagnetic amino acid in a peptide/protein in a step-by-step synthesis by employing nonsense suppressor methodology (e.g.. amber suppressor tRNA chemically aminoacylated with the desired amino acid) (Cornish et al. 1994; Mendel et al. 1995; Shafer et al. 2004) or solid-phase peptide synthesis (SPPS) (Merrifield 1963) based either on a Boc or Fmoc protection strategy. Although the first method might prove generally applicable in the future using unique transfer RNA(tRNA)/aminoacyl–tRNA-synthetase pairs (Chin et al. 2003), currently only few laboratories are equipped to apply the scheme successfully. On the other hand, modern improvement in peptide chemistry (Dawson and Kent 2000) allows SPPS of proteins containing as many as 166

MEMBRANE PROTEIN STRUCTURE AND DYNAMICS

Figure 1. (A) Reaction of the methanethiosulfonate (MTS) spin label with a sulfhydryl group, generating the spin label side chain R1. (B) Reaction scheme for the synthesis of a novel paramagnetic Boc-protected amino acid (Balog et al. 2004). Reaction conditions are: (a) 10% aq. NaOH, CH2Cl2, BuNHSO4; (b) 5% aq. H2SO4, EtOH; (c) Boc2O, THF; (d) 10% aq. NaOH, EtOH. (C) Mechanism of native and expressed protein ligation (IPL, EPL). The chemical ligation occurs between a C-terminal thioester and a sulfydryl group of an Nterminal Cys. After rearrangement through an N→S acyl shift a native peptide bond is formed. The reaction can be performed also in the presence of other unprotected cysteine residues because of a first reversible reaction and a second irreversible step. As an example, the paramagnetic amino acid presented in reaction scheme (B) has been incorporated in the fragment containing the N-terminal Cys. (D) Conformational space of MTS spin label attached to an α-helix. All flexible bonds within the R1 side chain are highlighted. Rotation around the last bond (–Cε–C1–) shows the highest flexibility (Beier and Steinhoff 2006, submitted for publication). (E) Conformational space of the novel spin label presented in reaction scheme (B) characterized by a short link between the protein backbone and the piperidine-oxyl moiety. (F) Conformational space of the rigid spin label TOAC attached to an α-helix. The flip of the six-membered ring as unique possible degree of freedom is shown in light gray representation.

3

4


amino acids (Becker et al. 2003; Kochendoerfer et al. 2003). Furthermore, combining solid-state synthesis with recombinant techniques can provide the tool for introducing unnatural amino acids at sites of choice in large proteins (Hahn and Muir 2005). 2.1. Spin Labeling of Cysteines A general scheme of the spin labeling procedure proven to yield good labeling efficiency is here briefly described and schematically depicted in Figure 1A. After cysteine substitution mutagenesis, the purified protein is usually stored in the presence of dithiothreitol (DTT) in order to prevent oxidation of the cysteine. Before spin labeling, the protein solution has to be dialyzed against the proper buffer to dilute the concentration of DTT. The protein (concentration adjusted to, e.g., 20 μM) is then usually incubated with 10-fold molar excess of spin label at 4°C overnight. The excess unbound spin label is afterward removed by dialysis, DEAE chromatography, Ni-NTA column (in the case the protein possesses a His-tag) or gel filtration. The spin-labeled protein is concentrated to 10–100 μM and filled into ESR quartz capillaries. At the X band (microwave frequency 9.5 GHz, magnetic field 0.34 T), the use of loop-gap or dielectric resonators provide the necessary sensitivity to yield continuous-wave ESR spectra with a good signal-to-noise ratio using 5 μL of sample within scan times of between 5 and 30 minutes. The spinlabeled cysteine/protein ratio is determined by double integration of the ESR spectra followed by comparison with standard solutions of MTS spin label and determination of the protein concentration. For interspin distance measurements (see §3.4) this ratio has to be close to one. The integrity and functionality of the spinlabeled protein variants have to be checked using independent methods prior to ESR analysis. 2.2. Incorporation of Spin-Labeled Amino Acids into Proteins For ESR applications a variety of paramagnetic α-, β-, and γ-amino acids have been synthesized. The paramagnetic protein building blocks can be incorporated into peptides or proteins via a step-by-step synthesis by employing nonsense suppressor methodology or Boc/Fmoc-based synthesis either on the solid phase (SPPS) or in solution. Among the paramagnetic α-amino acids, TOAC (4-amino-1-oxyl-2,2,6,6,tetramethyl-piperidine-4-carboxylic acid) (Rassat and Rey 1967) is by far the most popular, and it is characterized by only one degree of freedom (the conformation of the six-membered ring, see Figure 1F). It has been repetitively used to study the secondary structure of small peptides in liquid solution (Anderson et al. 1999; Bolin et al. 1999; Hanson et al. 1996a,b), and it has also been successfully incorporated into the α-melanocyte stimulating hormone without loss of function (Barbosa et al. 1999). However, a recent solid-state ESR analysis (Elsaber et al. 2005) of a series of peptides de novo designed to form an α-helix doubly labeled at different sites with TOAC demonstrates that the peptides assume a conformation not corre-


5

sponding to any of the common helical structures in proteins, highlighting one of the major problems arising by incorporating rigid unnatural spin-labeled amino acid into peptides or proteins. Several new N-Boc-protected paramagnetic amino acids with various side chains, having different polarity, orientation, and conformational restrictions, have been recently obtained using O´Donnell synthesis (Balog et al. 2003, 2004; Odonnell et al. 1989). The reaction scheme for the synthesis of a novel spin-label amino acid (Balog et al. 2004) (represented in its conformational space attached to an α-helix in Figure 1E) is presented as an example in Figure 1B. The chemical synthesis and semisynthesis of proteins with unnatural spinlabeled amino acids relies on an ability to produce the constituent peptides, typically by SPPS. The synthesis of peptides consisting of up to 60 amino acids incorporating a variety of biophysical probes is now routinely performed (Ayers et al. 1999; Kochendoerfer et al. 2004); however, incorporation of spin-labeled amino acids is still challenging in terms of identification of optimal synthesis conditions chemically compatible with the delicate nature of the nitroxide moiety that can be readily protonated under acidic conditions. This problem could be overcome in the case of TOAC by employing Fmoc-based synthesis strategies either on the solid phase in combination with HF cleavage or in solution (Marchetto et al. 1993; Monaco et al. 1999). In the case of pyrroline-based nitroxide amino acids, new strategies for their incorporation into peptides during SPPS have recently been successfully conducted. Additionally, coupling of a pyrroline nitroxide spin label to the ε-amino group of a lysine after complete assembly of the peptide has also been performed yielding a paramagnetic stable compound (Becker et al. 2005). Combining SPPS with recombinant techniques provides the tool for introducing paramagnetic amino acids potentially at every position in large proteins, and even membrane proteins. The EPL (expressed protein ligation, also named inteinmediated protein ligation IPL) technique can be used to semisynthesize proteins from recombinant and synthetic fragments, thus extending the size and complexity of the protein targets available to chemical engineering. The chemical ligation of two polypeptides requires that one of the fragments possess an N-terminal Cys and the other contain a C-terminal thioester moiety. After rearrangement through an S→N acyl shift, a native peptide bond is formed (Figure 1C). The reaction can be performed even in the presence of other unprotected cysteines. Using this strategy, a spin-labeled RBD protein has been recently synthesized, showing a stable paramagnetic center detected by ESR (Becker et al. 2005).

3. STRUCTURAL INFORMATION DERIVED FROM ESR SPECTRA ANALYSIS 3.1. Nitroxide Dynamics and Motional Freedom The sensitivity of the ESR spectra to the reorientational motion of the nitroxide side chain attached at specific sites in the protein has been extensively reviewed

6


(Berliner 1976, 1979; Berliner and Reuben 1989), and the relationship between side chain mobility and protein structure has been explored in detail for T4 lysozyme (Mchaourab et al. 1996). The motion of a nitroxide side chain is characterized by three correlation times: the rotational correlation time for the entire protein, the effective correlation time due to the rotational isomerization about the bonds linking the nitroxide to the backbone, and the effective correlation time for the segmental motion of the backbone with respect to the average protein structure. The first term can be often neglected when considering membrane-bound proteins or soluble proteins with molecular weight above 200 kDa. In these cases the rotational correlation time exceeds 60 ns, which is beyond the sensitivity of the ESR time scale. Estimation of the rotational correlation time for the overall tumbling can be obtained using the Stokes-Einstein equation:

τR =

ηV η VM = ⋅ , kBT kBT N A

(1)

where η is the viscosity of the solution, kB is the Boltzmann constant, T is the absolute temperature, and V is the volume of the protein, which can be written as the product of V , the partial specific volume, and M, the molecular mass, divided by Avogadro's number, NA. For most proteins dissolved in pure water at 293 K, V ≅ 0.73 g–1cm3 (Voet and Voet 2004). In the case of small-size proteins, the addition of sucrose (30% w/v) to the aqueous solution can be used to increase the viscosity by about a factor of three, thus minimizing the spectral effects due to overall tumbling of the molecule (Mchaourab et al. 1996; Qin et al. 2003). The effective correlation time due to the reorientation of the spin label is expected to be a complex function of the spin label molecular structure and the primary, secondary, tertiary, and eventually quaternary structure of the protein under investigation, while the effective correlation time due to backbone motion is related to backbone flexibility, and thus to the secondary structure of the protein. Generally the term “mobility” is used to define the effects on ESR spectral features due to the motional rate, amplitude, and anisotropy of the overall reorientational motion of the nitroxide label. Weak interaction between the nitroxide and the rest of the protein as found for helix surface sites or loop regions results in a high degree of mobility (Figure 2A, position 164). In this case the apparent hyperfine splitting and the line width of the ESR spectrum are as small as illustrated in Figure 2B. In turn, if the residual motion is restricted due to strong interaction of the nitroxide group with neighboring side chains or backbone atoms as found for tertiary contact or buried sites (Figure 2A, position 171), the apparent hyperfine splitting and the line width are increased (Figure 2B). In general, the resulting spectra cannot be simulated with a simple isotropic model of motion. Due to the interaction of the nitroxide with neighboring protein atoms, the motion has to be anisotropic, as was shown by molecular dynamics simulations (Steinhoff and Hubbell 1996; Steinhoff et al. 2000a). Additionally, a distribution of motional states can be concluded from those spectra that exhibit more than one component. In spite of the complex relation between nitroxide dynamics and spectral line width,


7

the line width of the central line (mI = 0), ΔH0, has been found to be correlated with the structure of the binding site environment and has been used as a simple semiempirical mobility parameter (Hubbell et al. 1996; Mchaourab et al. 1996). A second semi-empirical parameter that has been found to correlate with the spectral line width is the spectral breadth, represented by the second moment, 〈H2〉 :

∫ ( B − H ) S ( B)dB , = ∫ S ( B)dB 2

H

2

(2)

which is defined based on the first moment 〈H〉 (the mean or geometrical center of the spectrum): H =

∫ B S ( B)dB , ∫ S ( B)dB

(3)

where B is the magnetic field, and S(B) is the absorption spectrum of the spinlabeled protein. The numerical values of ΔH0 and 〈H2〉 increase for any particular motional geometry as the frequency of nitroxide reorientational motion is reduced. In order that these experimental parameters are increasing functions of mobility, the inverse of both quantities will be used as “mobility” parameters. At X-band frequencies both parameters depend on the degree of averaging of the anisotropic g and hyperfine tensors. In the presence of multicomponent spectra mobility parameter ΔH0–1 will be dominated by the most mobile component unless it is present in extremely small amounts; on the other hand, the second moment will be biased toward the less mobile component. The plot of these mobility parameters versus residue number reveals secondary structure elements through the periodic variation of mobility as the nitroxide sequentially samples surface, tertiary, or buried sites. The assignment of α-helices, βstrands, or random structures from the data is straightforward. A more quantitative interpretation of the experimental data in terms of dynamic mechanisms and local tertiary interaction requires simulation of the ESR spectra. Simulations can be performed on the basis of dynamic models developed by Freed and coworkers (Barnes et al. 1999; Borbat et al. 2001; Freed 1976) that show excellent agreement with the corresponding experimental spectra of solvated

8


Figure 2. (A) Ribbon representation of bacteriorhodopsin x-ray structure from Essen et al. (1998). The retinal chromophore is depicted in stick representation. Three positions subjected to sitedirected spin labeling are shown (A160R1, R164R1, F171R1). Side chain R1 (indicating the methanethiosulfonate spin label) is depicted in stick representation. The gray clouds define more than 90% of the complete accessible space of the nitroxide nitrogen as obtained from an MD simulation of 6 ns at 600 K, whereas the dark gray contours delimit areas with distinctly higher residence probability. Position 164 shows the largest cloud, indicating a higher degree of reorientation of the nitroxide ring during the MD simulation. The motion of the spin label attached at position 160 is slightly restricted due to interaction with the neighboring residues. On the other side the strongly restricted motion of the spin label attached at position 171, which is buried in the interior of the protein, is visualized by the small size of the cloud. (B) Comparison between simulated and experimental X-band ESR spectra for a set of positions in the BR (bacteriorhodopsin) loop region under investigation. The structural-based MD simulation approach used to simulate the spectra has been described in detail elsewhere (Beier and Steinhoff 2006, submitted for publication). The ESR spectra are depicted in the usual 1st derivative. The anisotropy of the hyperfine interaction is partially averaged due to the spin label motion for positions 158, 159, 162, and 164. The appearance of a prominent component characterized by higher line width and hyperfine extremes for position 160 is indicative for interaction of the spin label with neighboring residues restricting its reorientational motion. On the other side a powder-like spectrum results in the case of highly restricted motion (position 171), characterized by prominent positive and negative peaks at the extremes of the spectrum. (C) Plot of 〈H2〉–1 (inverse of the second moment) versus ΔH0–1 (inverse of the central line width) extracted from the experimental spectra of the set of positions investigated in the BR loop region. The black squares indicate the values extracted from the ESR spectra of trimeric BR in pur-


9

ple membrane lipids. The values obtained for several positions for the monomeric solubilized BR (in Triton X-100) are represented as open squares. The asterisks denote positions where quaternary interactions are present in the purple membrane samples and show distinct different spectra in the solubilized samples. The topological regions in the plot have been drowned according to (Isas et al. 2002; Mchaourab et al. 1996). The triangles connected by a dotted line represent mobility data for two positions characterized by high and low mobility. The effect of the variation of Azz from 3.4 (apolar microenvironment) to 3.7 (polar microenvironment) mT is shown (black: 3.4, dark gray: 3.5, light gray: 3.6, white: 3.7 mT). The mobility parameters are extracted from two spectra simulated using an isotropic Brownian diffusion model (Freed 1976). In the simulations all the parameters except Azz were kept constant. This allows to visualize independently the effect of a polarity change on the position of a point in the plot 〈H2〉–1 vs ΔH0–1. (D) Plot of the mean square fluctuation amplitude (variance) of Euler angle β calculated from the trajectories of an MD simulation of 6 ns at 600 K (Beier, unpublished) versus ΔH0–1 (inverse of the central line width) for the set of positions investigated in the BR loop region. The line width values are obtained from spectra from trimeric BR in purple membrane lipids and from solubilized monomeric BR (positions 158, 159, 162, and 164). The linear regression of the data is shown as a dotted line to highlight the correlation observed.

spin labels or spin-labeled lipids and proteins and give very valuable insight into the dynamics of these systems. On the other hand, simulations of ESR spectra can also be performed on the basis of molecular dynamics (MD) simulations. This approach facilitates the study of the influence of the structure in the vicinity of the spin label binding site on the ESR spectral line shape and allows one to verify, refine, or suggest structural models on the grounds of experimental ESR spectra of singly spin-labeled proteins (Steinhoff et al. 1996, 2000a). The strategy for detecting and interpreting the experimental data is illustrated with an example of structural details obtained for the E–F loop region of bacteriorhodopsin (BR). BR is a light-driven pump that translocates protons across the cell membrane of archaeal and halobacteria. Figure 2A shows the structure determined by x-ray diffraction. It consists of seven transmembrane helices that bury the chromophore retinal (Essen et al. 1998). The structure of the E–F loop was found to be only poorly resolved in several of the reported x-ray diffraction experiments, most probably due to structural disorder or its high flexibility (Luecke et al. 1998; Luecke et al. 1999; Pebay-Peyroula et al. 1997). A set of positions from the cytoplasmic end of helix-E, through the E–F loop to the cytoplasmic edge of helix-F (positions 157–171), have been subjected to site-directed spin labeling and investigated in terms of mobility of the attached spin label to define the structural properties of the E–F loop (Pfeiffer et al. 1999). Figure 2A shows three MTS spin labels bound at different positions in BR (160, 164, 171) and visualizes the motional behavior of the nitroxide group as determined from MD simulation (6 ns, 600 K). The gray clouds define more than 90% of the accessible space of the nitroxide nitrogen, whereas the black contours delimit areas with distinctly higher residence probability. A selection of experimental spectra obtained for trimeric BR complexes in purple membranes is shown in Figure 2B together with the simulated spectra obtained from the molecular dynamics simulation approach based on the BR monomeric structure determined by Essen et al. (1998) (PDB code entry 1BRR). The

10


agreement between experimental and simulated spectra shows how the MD simulation performed on protein sites of known x-ray structure can be a useful tool to study in detail the influence of the protein microenvironment in the vicinity of the spin label binding site on the ESR spectral line shape. A detailed analysis of the quality of the simulation with respect to the experimental ESR data can facilitate refinement of dynamic domains that are usually not properly defined in x-ray structures (as in the case of the E–F loop in BR) or the modeling of unknown structural domains when x-ray data are not available. The correlation between the inverse of the second moment, 〈H2〉 –1, and the inverse of the central line width, ΔH0–1, extracted from the experimental data is presented in Figure 2C. Side chains from different topographical regions of the protein can be classified on the basis of the x-ray structures of T4 lysozyme and annexin 12 (Hubbell et al. 1996; Isas et al. 2002; Mchaourab et al. 1996). This general classification of the regions of the plot accommodating buried, surface-exposed, or loop residues is presented in Figure 2C. Since this coarse classification is mainly derived from α-helical structures, subdivisions might be necessary — for example, when investigating β-sheet structures, random coil, or coiled-coil structures. It is interesting to observe that the ability to distinguish among different topological classes is better expressed in the line width rather than in the second moment. For example, a value for 〈H2〉–1 of 0.38 mT–2 spans three groups in the plot of Figure 2C, namely the buried, helix/contact, and loop/contact, which are on the contrary well separated when considering the line width parameter. In addition to the mobility correlation due to the dependence on the degree of averaging of the anisotropic hyperfine and g tensors, the line width and second moment are also sensitive to the value of the Azz hyperfine tensor component, which is known to vary for spin labels sampling different water/protein/lipid interfaces (see §3.3). The effect of the variation of Azz from 3.4 (apolar microenvironment) to 3.7 (polar microenvironment) mT is shown in Figure 2C for two sites characterized by high and low mobility. The mobility parameters are extracted from two spectra simulated using an isotropic Brownian diffusion model (Freed 1976). In these simulations all the parameters except Azz were kept constant. This allows to visualize independently the effect of polarity changes on the position of a point in the plot of 〈H2〉–1 versus ΔH0–1. It is evident that the higher the mobility, the more relevant are the variations induced by polarity changes on both mobility parameters. In fact, increasing the Azz value from 3.4 to 3.7 mT decreases 〈H2〉–1 by 14 or 12% and decreases ΔH0–1 by 10 or 5% for the mobile or restricted sites, respectively. The second moment is found to be more sensitive than the line width to the polarity variations due to the fact that Azz changes mainly influence the outer extremes of the ESR spectra. The investigated positions of BR are found in different regions of the plot presented in Figure 2C, ranging from buried to loop/surface topology. Evidence from solubilization experiments suggests that the spectra obtained in the purple membrane samples for positions 157, 161, 168, and 169 are biased by the contribution of quaternary contacts between BR molecules in the trimeric complex (Beier and Steinhoff 2006, submitted for publication). Despite the valuable information ob-


11

tained from the ESR spectra on the quaternary structure, the correlation between the mobility parameters and the structural properties of the monomeric BR looses its relevance. To overcome this problem, the mobility parameters obtained from detergent-solubilized monomeric BR are analyzed for the affected positions. The highest mobility is observed for positions 161 and 162, which are located in the loop/surface region of the plot. The following positions, 164 and 165, characterized by slightly restricted motion, are located in the loop/contact region of the plot. The preceding (158, 159) and following (166, 167) positions are characterized by mobility values, indicating tertiary contacts with neighboring residues (helix/contact region of the plot). Interestingly, restricted motion dominates the spectra of positions 160 and 163, which are in proximity to the most mobile sites. This is strong evidence for a loop conformation with these two nitroxide side chains oriented toward the protein. The mobility data suggest an arrangement of residues 160 to 165 within a single turned loop (Pfeiffer et al. 1999). Molecular dynamics studies performed on site-directed spin-labeled proteins provide the means to extract detailed information on the reorientation dynamics of the nitroxide ring and its interaction with the microenvironment (LaConte et al. 2002; Steinhoff and Karim 1993). From an analysis of the transformation matrix, defined in terms of three Eulerian angles α, β, γ, rotating the nitroxide coordinate system x, y, z (shown in Figure 4) into the laboratory coordinate system (X, Y, Z), the reorientation of the nitroxide ring with respect to the direction of the external B field (Z axis) can be inferred. The variance of angle β during a 6-ns MD simulation at 600 K for a set of positions in BR is plotted versus the experimentally determined ΔH0–1 values in Figure 2D. Although the variance of β does not reflect the complexity of the anisotropic reorientational motion of the nitroxide, the clear correlation revealed in Figure 2D proves once more that the information on the local protein structure is coded in the ESR line width. The set of ESR spectra simulated using the MD simulation approach based on the structural model obtained from x-ray diffraction data by Essen and coworkers (1998) nicely agrees with the experimental data. This result validates this x-ray structural model and provides evidence that the loop structure found under physiological conditions is retained in the crystal state. 3.2. Solvent Accessibility of the Attached Nitroxides

The motional analysis of elements of secondary and tertiary structure can be supplemented by measuring the accessibility of the nitroxide toward a paramagnetic probe selectively partitioning in different environments to define the topology of the natural side chain with respect to the protein/water/membrane boundaries. By definition, the “accessibility” of R1 in a macromolecule is proportional to the Heisenberg exchange frequency (Wex) of the nitroxide with a paramagnetic exchange reagent present in solution. The paramagnetic exchange reagents employed in the measurements of accessibility may be metal ion complexes, molecular oxygen, organic radicals, and even a second nitroxide. Polar metal complexes (Ni(II)

12


complexes or chromium oxalate (CrOx)) preferentially partition into the aqueous phase, whereas the product of concentration and diffusion coefficient of apolar oxygen has been shown to increase with distance from the membrane/water interface and exhibits a maximum value in the center of the membrane bilayer (Altenbach et al. 1994; Marsh et al. 2006). Due to the fact that the Heisenberg exchange mechanism leads to variations in the relaxation properties of the spin label attached at the investigated sites, each method capable to determine the variation of T1e (spin–lattice relaxation time) or T2e (spin–spin relaxation time) of a spin label bound to a protein upon addition of the paramagnetic exchange reagent is in principle suitable to unravel the topology of the nitroxide with respect to different interfaces. In general, if the longitudinal relaxation time of the reagent is smaller than the encounter complex lifetime (T1R < τc), the Heisenberg exchange leads to equal changes in T1e and T2e: ⎛1⎞ ⎛1⎞ Wex = Δ ⎜⎜⎜ ⎟⎟⎟ = Δ ⎜⎜⎜ ⎟⎟⎟ . ⎜⎝ T1e ⎠⎟ ⎝⎜ T2e ⎠⎟

(4)

For a bimolecular encounter between a small nitroxide moiety in solution (N) and an exchange reagent (R) the Heisenberg exchange frequency experienced by the nitroxide is Wex = kex CR ,

(5)

where kex is the exchange constant and CR is the concentration of the reagent (R). In the strong exchange limit, when Heisenberg exchange is diffusion controlled, kex (expressed in L mol–1 s–1 unit) can be written as (Altenbach et al. 2005) kex = Pmax fkD = Pmax f ( 4π N A ( DN + DR )rC ⋅103 ) ,

(6)

where Pmax is the maximum exchange efficiency (Pmax = 1 if the conditions of strong exchange limit and T1R < τc are fulfilled), f is the dimensionless steric factor, kD is the diffusion-controlled rate constant (expressed in L mol–1 s–1), NA is Avogadro's number, D is the diffusion constant for the nitroxide (N) and reagent (R) (expressed in m2 s–1), and rC is the collision radius (expressed in m, the sum of the effective radii of N and R). In the case of a spin label bound to a protein, DN becomes the diffusion constant of the protein, which is small compared to DR and can be neglected in the equation. Effects induced by the local protein environment and the interactions with neighboring residues can reduce the effective collision frequency of the nitroxide toward the reagent. In this case Wex can be expressed as

Wex = ρf 4π

NA DN rC CR , 1000

(7)

where ρ is a proportionality factor, which takes into consideration the above mentioned effects due to the presence of the protein (Altenbach et al. 2005). Among the possible methods to determine Wex, those based on T1e determination are the most common. Multiquantum ESR is a recently developed method that could be exploited to measure relative accessibilities in proteins (Mchaourab and


13

Hyde 1993). ELDOR (electron electron double resonance) can also be employed to measure absolute values of Wex when the exchange reagent is a second nitroxide (Feix et al. 1984; Shin and Hubbell 1992). Saturation recovery ESR (SR-ESR) provides the means to measure T1e directly as well as Wex (Altenbach et al. 1989b; Nielsen et al. 2004; Percival and Hyde 1975; Yin et al. 1987), and it has been recently applied to the study of T4 lysozyme (Pyka et al. 2005). Its applicability might be of relevance to identify multiple states of the side chain that are usually visible in the cw ESR spectra and to characterize them in terms of relaxation properties. The cw power saturation technique has been traditionally the technique of choice to estimate relaxation rates through the calculation of accessibility parameter Π in the presence of paramagnetic reagents soluble in different environments. Determination of Π in combination with nitroxide scanning has been shown to be a valuable tool in obtaining structural information for soluble and membrane proteins (Cuello et al. 2004; Hubbell et al. 1998). In this case the cw ESR signal height is monitored as a function of incident microwave power. The observed rise to a maximum value and the following decrease of the signal vs. the square root of the incident power gives the so-called saturation curve, from which a parameter (P1/2) related to the relaxation rates of the nitroxide can be determined (Figure 3B). In the following a brief description of the theoretical background of cw saturation method and its application is presented. For a single homogeneous Lorentzian line the peak-to-peak amplitude of the first derivative absorption spectrum, Y′, has the form (Poole 1983)

Y′ ∝

H1

(1 + H

2 1

γ T1eT2e )

3/ 2

2

=

Λ P

(1 + Λ

2

P γ 2T1eT2e )

3/ 2

,

(8)

where H 1 is the microwave magnetic field component in Gauss (which is proportional to the square root of the microwave power. P, with Λ depending on the properties of the resonator), γ is the gyromagnetic ratio of the electron, and T1e and T2e are the spin–lattice and spin–spin relaxation times, respectively. From the above equation it is clear that Y′ is proportional to√P for low microwave (MW) power and to 1/P for higher MW power. Saturation occurs when the amplitude of the signal deviates from the square root dependency on power, and it is due to the inability of the spin system to relax to equilibrium at a rate competitive with that at which power is absorbed. The relaxation time is measured by T1e. An easily extractable parameter related to the relaxation quantity T1eT2e is P1/2, defined as the MW power at which the signal amplitude reaches half of the theoretical value reachable in the absence of saturation: Y ′ ( P1/ 2 ) =

Λ P1/ 2 2

.

(9)

14


Figure 3. (A) Cartoon representation of NpSRII/NpHtrII 2:2 complex embedded in the lipid bilayer. The ribbon represents the x-ray structure (PDB code entry 1H2S) of the transmembrane domain of the complex; a schematic representation of the predicted α-helical AS-1 sequence protruding perpendicularly to the membrane plane into the cytoplasmic phase is also depicted (positions 83–101). For the sake of clarity, the AS-1 cylinder is proportionally stretched according to the length of the sequence. Two putative helix-breaking residues (G83 and G84) are represented as the gray lines connecting the helices TM2 to the AS-1 cylinders. The black and gray arrows indicate the gradient of concentration of the water-soluble reagents (CrOx and NiEDDA) and the lipid-soluble reagent (O2), respectively. (B) Plot of the saturation curves (peak-to-peak intensity of the central ESR line (Y′) vs. the square root of microwave power) obtained for position 96 in the transducer in NpSRII/NpHtrII complexes reconstituted in purple membrane lipids. The curves are obtained in the presence of N2, air, and CrOx 50 mM. The P1/2 values obtained by fitting of the data are 7.5, 15.5, and 95 mW, respectively (the estimated error is 10%). The accuracy in determination of the P1/2 values decreases for very accessible sites. In these cases it is recommended to repeat the saturation experiment, decreasing the concentration of the reagent and subsequently scaling the Π values obtained to the usual concentration of the reagent. (C) Accessibility parameters ΠCrox (black circles) and ΠO2 (gray circles) calculated according to Eq. (15). The Πair values obtained experimentally were multiplied by 5 to scale them to 100% O2. The open circles in the same color code represent the values calculated according to Eq. (14) for positions 88–96. Deviations in the Π values due to changes in the line width are negligible in the presence of air through all the positions observed but become evident in the presence of CrOx 50 mM for the most mobile positions (92–96). The dark gray line highlights the 3.6 periodicity in ΠO2 values.


15

The expression of P1/2 for a homogeneous line can then be written as P1/ 2 =

(22 / 3 −1) Λ 2 γ 2T1eT2e

.

(10)

Rewriting Eq. (8) accordingly, one obtains Y′ ∝

Λ P

(1 + (2

ε

ε

−1) P / P1/ 2 )

,

(11)

with ε = 3/2. An analogous expression can be obtained in the inhomogeneous saturation limits with ε = 1/2. To fit the experimental data of Y´ vs. √P, the following equation is used: Y′=

I P

(1 + (2

ε

ε

−1) P / P1/ 2 )

,

(12)

with I, ε, and P1/2 as adjustable parameters (Figure 3B). In the presence of exchange paramagnetic reagents colliding with the nitroxide, variations in the relaxation times of the nitroxide are reflected in changes of P1/2 values. Thus, calculating the P1/2 values in the presence and absence of reagents leads to an estimation of the Heisenberg exchange frequency, which in turn is proportional to the topological properties of the nitroxide bound to a protein: ΔP1/ 2 = P1/R2 − P1/0 2 ∝

W 1 1 − ∝ ex0 . T1eRT2eR T1e0T2e0 T2e

(13)

The equation has been approximated due to the fact that being T2e 10 mM, one can decrease the concentration of the exchange reagent until the effect on the line width is negli-

16


gible. An alternative approach consists in dividing P1/2R and P1/20 by ΔH0R and ΔH00, respectively, to eliminate the T2 dependency of ΔP1/2: ΔP1/′ 2 =

ΔP1/R2 ΔP1/0 2 − . ΔH 0R ΔH 00

(15)

To account for instrumental variability, normalization of this quantity with a standard reference (std) such as a dilute 2,2-diphenyl-1-picrylhydrazyl (DPPH) powder in KCl leads to dimensionless parameter Π: Π = ΔP1/′ 2

ΔH (std) ∝ Wex . P1/ 2 (std)

(16)

The deviations in Π values calculated utilizing the two formulas for ΔP1′/ 2 (Eqs. (14) and (15)) are negligible for nitroxides characterized by restricted motional behavior but become significant for very mobile nitroxides (Figure 3C). In general, the use of air (21% O2) and NiEDDA (10 mM) as exchange reagents to monitor the accessibility of the bound nitroxide toward lipid and water phase, respectively, fulfills the T2eR ≅ T2e0 criterion. The neutral NiEDDA is a more suitable exchange reagent with respect to CrOx (negatively charged) to monitor accessibility of lipid/water interfacial region. Contribution of Coulomb interactions might in fact lead to variations of the accessibility parameter in the presence of charged membrane surfaces or protein segments and even high salt concentrations (Nielsen et al. 2004). Experimentally, to obtain a cw power saturation curve and extract the Π values, the sample is loaded into gas-permeable TPX capillaries. Loop-gap resonators are preferentially used that allow sample volumes less than 10 μL and provide a homogeneous B1. The samples are first deoxygenated for 20 minutes by fluxing nitrogen around the sample capillary. The saturation curve is then detected in the presence of N2. Subsequently, for oxygen accessibility experiments, nitrogen is replaced by air (21% oxygen) or 100% O2. For CrOx or NiEDDA accessibility experiments a negligible volume of a concentrated stock solution is added to the sample to obtain the desired final concentration. The peak-to-peak amplitude of the central ESR line is measured at different incident microwave power levels and plotted vs. √P (Figure 3B). To extract the P1/2 values, homemade fitting programs can be used. As an example, the accessibility analysis performed on a segment of the transducer protein (HtrII) associated with the photophobic receptor sensory rhodopsin II (SRII) from Natronomonas pharaonis (Np) reconstituted in purple membrane lipids is presented (Bordignon et al. 2005) (Figure 3). It has been shown that the NpSRII/NpHtrII (Natronomonas pharaonis sensory rhodopsin II/Natronomonas pharaonis halobacterial transducer II) complex is present in a 2:2 stoichiometry in membrane, and x-ray data are available for NpSRII in complex with the transmembrane region of NpHtrII up to residue 82. Following the transmembrane region, constituted by two helices (TM1 and TM2), the transducer contains a linker mem-


17

brane adjacent region (starting with a conserved signal transduction domain called HAMP (Appleman et al. 2003)) and a long (∼2 nm) cytoplasmic signaling domain. Light activation of NpSRII leads to a displacement of helix-F, which in turn triggers a rotation/screw-like motion of TM2 in NpHtrII. This conformational change is thought to be transmitted through the HAMP domain to the cytoplasmic signaling domain of the transducer. The architecture and function of the HAMP domain are still unknown. Based on primary sequence homologies, it is predicted to consist of two helical amphipathic sequences (AS-1 and AS-2) connected by a sequence of undefined secondary structure. The accessibility data obtained for 20 spin-labeled samples with the spin labels located at the end of the transmembrane part of TM2 (positions 78–82, for which x-ray data are available) and in the first part of the AS-1 sequence of the HAMP domain (83–96, for which no structural information is available) are here described and discussed in terms of structural and topological information obtainable. A set of saturation curves obtained in the presence of N2, air, and 50 mM CrOx for position 96 is presented as an example in Figure 3B. From a comparison of the saturation curves, it is apparent that the highest accessibility is toward CrOx (the maximum of the curve is not reached at the maximum power measured). The calculated Π values are depicted as a function of sequence position in Figure 3C. For better visualization, the Πair values have been multiplied by 5 to scale them to 100% O2, according to their proportionality to the exchange reagent concentration. The low Π values found for both CrOx and O2 for side chains 78–86 indicate residues in a densely packed protein–protein interface. The low mobility values measured for the same region (data not shown) corroborate this observation. The clear periodicity of 3.6 residues revealed by the oxygen accessibility from positions 78 to 89 strongly supports the idea that the first part of the AS-1 sequence is the α-helical extension of TM2. The statistical relevance of the accessibility properties in a nitroxide scanning experiment is a useful tool for structural analysis (Hubbell et al. 1998). Interestingly, higher oxygen accessibility values are observed for the positions oriented toward NpSRII, indicating a dense packing of residues in the transducer–transducer interface. For positions 87 to 94, the ΠCrox values gradually increase in line with the notion that the side chains become more exposed to the bulk water (a gradual increase in the spin label mobility is also observed). This finding provides strong evidence that region 87–94 of the transducer is protruding out of the protein–protein interface into the cytoplasmic phase. Accordingly, the ΠO2 values are also increasing. The faster increase of O2 than CrOx accessibility might be ascribed both to the smaller steric hindrance of the oxygen molecule, which can easily penetrate the interfacial regions, and to the presence of the charged membrane bilayer surface that might interfere with the CrOx diffusion. The periodicity observed in this region cannot be directly ascribed to a helical pattern, although it might be present. The reason for this failure to recognize possible secondary structure elements might be the fact that for interfacial regions the accessibility/mobility data lose their statistical significance (Cuello et al. 2004).

18


Positions 95 and 96 show ΠCrox values approaching those typical for waterexposed residues. The almost constant level of ΠO2 (∼0.6) is also in line with residues fully exposed to water. The high mobility observed suggests a very dynamic region lacking fixed structural features. The results obtained from the accessibility analysis show the potentiality of this method, capable to validate the x-ray structural data available so far up to position 82 (closely packed α-helix) and defining the secondary and tertiary structure of the following region, where no structural information was previously available (83–89 α-helix; 86–94 first segment protruding in the water phase; 95–96 waterexposed residues). 3.3. Polarity of the Nitroxide Micro-Environment

Characterizing certain regions of the proteins according to their polarity and proticity profiles or identifying hydrophobic barriers along ion channels in transmembrane proteins may provide the means to obtain structural and topological details of proteins and to understand specific biological processes on the molecular level. The observed dependencies of hyperfine component Azz and g tensor component gxx to the polarity and proticity of the spin label microenvironment make SDSL ESR a useful technique in this respect. Generally, a polar environment shifts tensor component Azz to higher values, whereas tensor component gxx is decreased. While the Azz component can be easily obtained from cw X-band ESR spectra of spin-labeled proteins in frozen samples, the principal g-tensor components and their variation due to solute–solvent interactions can be determined with high accuracy using high-field ESR techniques due to the enhanced Zeeman resolution of rigid-limit spectra of disordered samples (Burghaus et al. 1992; Huber and Törring 1995; Prisner et al. 1995). In a sequence of a regular secondary structure with anisotropic solvation, the water density and hence the tensor component values are a periodic function of residue number and can be used similarly to the accessibility values for water soluble paramagnetic reagents (see §3.2) to extract structural and topological information on the protein under investigation. Moreover, the polarity of the protein environment surrounding the label may reveal detailed information on protein folding. Theoretically, both gxx and Azz are expected to be linearly dependent on ρπO, the π spin density on the oxygen atom of the nitroxide group. The hyperfine component Azz, given by (Plato et al. 2002), Azz = QπN ρπN

(17)

is linearly dependent on ρπO on account of sum condition ρOπ + ρπN ≅ 1 ,

(18)

which follows from the fact that the spin density is practically fully confined to the nitroxide group. For gxx apart from a direct proportionality to ρπO, there is an addi-


19

tional dependence on the specific properties of the oxygen lone-pair orbitals. This originates from an approximate expression for g xx derived for organic π-radicals (Stone 1963): g xx ≅ g e +

2ς (O)ρOπ cny2 ΔE n π *

,

(19)

where ge = 2.0023 is the free-electron g value; ζ(O) is the oxygen spin–orbit coupling parameter; ρπO is the π spin density on the oxygen 2pz atomic orbital; cny is the LCAO coefficient of the 2py atomic orbital contributing to the oxygen lone-pair MO; and ΔEnπ* = Eπ* - En is the n → π* excitation energy. The polarity dependence of ρπO follows from the variation of ρπN due to the interaction between the permanent electric dipole induced by the charge displacement in the NO π bond and the various intermolecular fields in the NO bond region (which can be described by an average local field E, Ex is the electric field component along the NO bond) (Griffith et al. 1974; Plato et al. 2002): ΔρπN = C1 Ex

(C1 > 0) .

(20)

Moreover, the lone-pair orbital energy En is known to be sensitive to the electrostatics, i.e., the polarity of the environment, and particularly sensitive to H bonding of the lone-pairs to water or to polar amino acid residues in the surrounding medium. Furthermore, H bonding can also affect the partial electron population cny2 of the lone-pair orbitals. Thus any change in δΔgxx = gx x - ge can be ascribed to three contributions: 2 δΔg xx ρπN δAzz δΔEnπ* δcny ≅− O − + 2 . Δg xx ρπ Azz ΔEn π * cny

(21)

Qualitatively, Eq. (21) correctly predicts the observed negative slope of the gxx vs. Azz plot (Möbius et al. 2005; Plato et al. 2002; Steinhoff et al. 2000b; Wegener et al. 2001b). The last two terms account for additional vertical displacement. As an example, the 95-GHz high-field ESR analysis of the polarity profile along the putative proton channel of bacteriorhodopsin (BR) is presented and discussed (Figure 4). For the purpose of polarity analysis, W-band ESR spectra have to be recorded at low temperatures to avoid motional averaging of the anisotropic tensors. In the temperature regime below 200 K the reorientational correlation time of an otherwise unrestricted spin-label side chain exceeds 100 ns (Steinhoff et al. 1989), i.e., the nitroxide may be considered as immobilized on the ESR time scale. Librational motion of small amplitude still prevails and may lead to small deviations of the measured tensor values from their rigid limit values due to partial motional averaging. However, this effect is small (e.g., less than 2% for Azz at 200 K (Steinhoff et al. 1989)) and is further minimized by decreasing the temperature for data collection to 120 K. Spectra of selected BR mutants spin labeled with MTS in

20


Figure 4. (A) Experimental high-field ESR spectra (1st derivative, T = 200 K, 95 GHz) for a set of spin-labeled BR variants (the Cα atoms of the chosen residues are indicated as a ball representation in (B)). The hyperfine splitting in the gxx and gyy regions is not completely resolved. The vertical line marks the gxx position of T46R1. The variation of gxx reflects the change of polarity in the microenvironment of the spin label when moving through the BR proton channel from the cytoplasmic to the extracellular side. The molecular structure of the “probe head” of the spin label MTS is depicted on top, together with the x, y, z coordinate system in use. (B) Structural model of BR. The chromophore retinal and residues D96 and D85 participating in proton transfer are indicated in stick representation. (C) The magnitude of tensor elements Azz and gxx of the spin labels are plotted as a function of the nitroxide location in the protein with respect to position 129 (see (B)). The Azz values are determined by fitting of the X-band spectra detected at 170 K (Steinhoff et al. 1999); the gxx values are extracted from the W-band spectra detected at 200 K (Steinhoff et al. 2000b). The vertical bars indicate the range of the possible location of the NO group (0.7 nm); the horizontal bars indicate 2σ errors for Azz and gxx of about ±10–2 mT and ±2 × 10–5, respectively. The dotted horizontal lines indicate the positions of the proton donor D96, retinal, and D85. The values of gxx and Azz can be treated as a polarity index, and thus the plotted data reflect the hydrophobic barrier the proton has to overcome on its way through the BR proton channel.

the proton channel are shown in Figure 4A. The variation of g xx with the nitroxide binding site is clearly revealed by the shift of the position of the low-field maximum. The variations of gxx and Azz with the nitroxide binding site can be measured with high precision. Plots of gxx and Azz vs. nitroxide position along the proton channel reveal a characteristic variation in the polarity of the nitroxide microenvironment (Figure 4C). Residue S162 is located in the E–F loop at the cytoplasmic surface (for the residue positions see Figure 4B), whereas residue K129 is positioned in the D–E loop on the extracellular surface. The high polarity in the environment of these residues (Azz ≈ 3.7 mT, gxx ≈ 2.0083) is clear evidence that the nitroxides are accessible to water, which is in agreement with the structural details.


21

The environmental polarity of the nitroxide at positions 100, 167, and 171 is significantly less and reaches its minimum at position 46 between proton donor D96 and the retinal. The plot directly reflects the hydrophobic barrier that the proton has to overcome on its way through the proton channel. The question if the structural changes occurring in BR upon illumination are a prerequisite for vectorial proton transport and how they modify this hydrophobic barrier remains to be elucidated. SDSL together with high-field ESR might provide the tools to unravel this issue. Together with the polarity analysis discussed above, characterization of the microenvironment of the spin-label side chain in terms of protic and aprotic properties (Möbius et al. 2005; Steinhoff et al. 2000b) can be obtained from an analysis of the plot gxx vs. Azz. Furthermore, the polarity variation with bilayer depth across the lipid-exposed surface of a transmembrane α-helix might be used to locate a spin-labeled site with respect to the lipid–water interface, thus providing valuable information on the orientation of membrane-associated proteins or membrane anchors. 3.4. Inter-Nitroxide Distance Measurements

The simultaneous exchange of two native amino acids by cysteines followed by modification with MTS spin labels allows determination of inter-residual distances and thus provides a strategy for deducing proximity of selected secondary structural elements (for reviews see Berliner et al. 2000; Hustedt and Beth 1999). Together with mobility, accessibility, and polarity analysis, interspin distance determination provides a unique means for deducing structural details in proteins of unknown structure. The spin–spin interaction between two spin labels attached to a protein is composed of static dipolar interaction, modulation of the dipolar interaction by the residual motion of the spin-label side chains, and exchange interaction. The static dipolar interaction leads to considerable broadening of the cw (continuous wave) ESR spectrum if the interspin distance is less than 2 nm. For unique orientations of the nitroxides relative to each other as found for spin labels introduced at buried sites in the rigid limit, a rigorous solution of the spin Hamiltonian of the system can be obtained. Spectra simulations yield the distance between the nitroxides and the Euler angles describing their relative orientation and that of the interspin vector relative to the magnetic field (Hustedt et al. 1997; Hustedt et al. 1999). However, the most frequently encountered case in SDSL consists of nitroxides that adopt a statistical distribution of distances and relative orientations. Values of the interspin distances can be determined from a detailed line shape analysis of spectra measured below 200 K (Rabenstein and Shin 1995; Steinhoff et al. 1991; Steinhoff et al. 1997) or in solutions of high viscosity (Altenbach et al. 2001) using spectra convolution or deconvolution techniques. The ESR spectra simulation program DIPFIT (Steinhoff et al. 1991, 1997), e.g., employs automated routines to determine best-fit parameters and considers a Gaussian distribution of interspin distances and variable contributions of singly spin-labeled protein.

22


Since the line width of the spectra is a steep function of the interspin distance (see Figure 5A, upper spectra), empirical or semi-empirical parameters as spectral amplitude ratios (Likhtenshtein 1976) or spectral second moment values are often sufficient to answer structural questions (Mchaourab et al. 2000; Radzwill et al. 2001). The lower limit for reliable distance determination using the above methods is given by the increasing influence of exchange interaction with decreasing interspin distances due to partial overlap of the nitrogen π-orbitals of the two interacting nitroxides. This effect makes quantification of interspin distances of less than 0.8 nm difficult (Closs et al. 1992; Fiori and Millhauser 1995). A simplified quantitative interspin distance determination can be performed using the difference between the spectral second moments (see Eq. (2)) calculated for the doubly and the singly labeled protein:

∫ ( B − H ) S ( B)dB − ∫ ( B − H ) S ( B)dB , ∫ S ( B)dB ∫ S ( B)dB 2

ΔH 2 =

2

D

D

S

(22)

S

where Si is the absorption spectrum of the doubly spin-labeled protein sample (I = D) or the spectrum without spin–spin interaction determined from the superposition of the two corresponding singly spin-labeled protein samples (I = S). This difference does not depend on exchange interaction and is less sensitive to incomplete spin labeling in comparison to the analysis of spectral amplitude ratios. In the case of a radical pair, the relation between the difference of second moments and the interspin distance, r, yields:

r = 2.32 ΔH 2 ⋅108

−1

6

nm ,

(23)

with 〈ΔH2〉 given in T2. The upper limit for distance determination using the method of second moments is between 1.5 and 1.7 nm, depending strongly on the quality of the baseline. Convolution and deconvolution techniques were shown to be applicable for inter-nitroxide distances up to 2 nm and provide valuable information on distance distributions. Recent advances in pulsed ESR techniques lead to protocols that expand this range from 2 to 8 nm (Borbat and Freed 1999; Pannier et al. 2000). Hence, cw ESR and pulsed ESR methods perfectly complement each other and provide interspin distances in the range between 0.8 and 8 nm. The method of interspin distance determination has been successfully applied to a number of proteins, including rhodopsin (Altenbach et al. 1996; Cai et al. 1997), lactose permease (Voss et al. 1998), the KcsA potassium channel (Gross et al. 1999; Perozo et al. 1998), and alpha-crystallin (Koteiche et al. 1998, 1999). These applications demonstrate the capability and reliability of the SDSL method in resolving protein structure at the level of the backbone fold. As an example for the elucidation of protein–protein interaction using dipoledipole interaction, the structural model of the transmembrane region of the complex consisting of the photophobic receptor sensory rhodopsin II from N. pharaonis (NpSRII) and its transducer (NpHtrII) is presented (already introduced


23

in §3.2). The model derived from ESR data (Wegener et al. 2001a) is shown in Figure 5B and compared to the x-ray structure subsequently released for the transmembrane part of the complex (Gordeliy et al. 2002) (depicted in light grayscale in Figure 5C). The x-ray data for the transducer, however, are still incomplete, being that only the transmembrane region was successfully crystallized. The crystal structure ends in fact at position 82 in transducer helix TM2, closer to the membrane–water interface, leaving the rest of the protein (which consists of more than 500 amino acids) completely undetermined. Primary sequence analogies suggest that the long cytoplasmic tail of the transducer shows similarities with the coiledcoil helical bundle of the chemoreceptors from E. coli (Kim et al. 1999); however, the structure of the membrane adjacent region of the protein, which is thought to be the signal transduction domain, is completely undetermined. To obtain information on this region of the transducer, a nitroxide scanning starting from the last position in TM2 for which x-ray data are available (L82) has been performed. The recent structural model derived from the ESR data for the region of the transducer beyond position 82 is also briefly presented (Figures 5C–D, dark grayscale) (Bordignon et al. 2005). The structural analysis of the transmembrane part of the complex was based on interspin distances for 26 pairs of spin labels introduced into transmembrane helices TM1 or TM2 of the transducer and helices F or G of the receptor. Considerable spin–spin interactions observed for the singly labeled transducer with the nitroxide side chain located at position 78 (Figure 5A) or 82 suggested a dimeric arrangement of TM2 and TM2′. The strongest interaction corresponding to the closest distances (less than 1.2 nm) between NpSRII and NpHtrII are observed between positions 25TM1–210G, 81TM2–210G, as well as between 81TM2–211G, indicating that both TM1 and TM2 are close to helix G. Distances in the range from 1.2 to 1.6 are determined for the following five TM2-NpSRII pairs: 80TM2–158F, 80TM2–157F, 80TM2–210G, 80TM2–211G, and 81TM2–158F. According to these interspin distance values, TM2 has to be located between C-terminal helices F and G. The dimer formation of the NpSRII/NpHtrII complex in membranes is suggested to orient the respective positions 22 of TM1 and TM1′ toward each other. These findings and the distance values determined between the spin-labeled positions of TM1, 21, 23, 24, and 25 with 210G or 158F are in agreement with this arrangement. These results revealed a quaternary complex between two copies of NpHtrII and NpSRII each (2:2 complex), with apparent twofold symmetry. The model proposed based on the ESR data is presented in Figure 5B. The structure of NpSRII is based on a crystal structure of bacteriorhodopsin (Essen et al. 1998) validated for the investigated positions in NpSRII. Transducer TM1 and TM2 helices are modeled as canonical α-helices. The arrangement of NpSRII and NpHtrII proposed based on ESR distance determination (Wegener et al. 2001a) is in agreement with the one found in the crystal (Figure 5C, light grayscale) (Gordeliy et al. 2002; Klare et al. 2004a), proving the reliability of the SDSL ESR methodology to unravel structural details of proteins. Moreover, the agreement between the model and the x-ray structure proves that arrangement of the transmembrane part of the complex observed in

24


Figure 5. (A) The upper spectra marked with an asterisk are reference powder spectra simulated with the DIPFIT program. The parameters used are gxx = 2.0086, gyy = 2.0066, gzz = 2.0026, Axx = 0.52 mT, Ayy = 0.45 mT, and Azz= 3.5 mT. The spectra were convoluted with a field-independent line shape function composed of a superposition of 44% Lorentzian and 56% Gaussian of 0.3 and 0.39 mT width, respectively. The fraction of the singly spin-labeled component has been chosen to be 30%. Clearly, the lower the amount of singly labeled protein present in the sample, the more pronounced the spectral changes due to dipolar broadening appear. To visualize the decrease of the central line amplitude in the spin-normalized spectra due to dipolar interaction, interspin distances of >2 (gray line), 1.8, 1.4 and 1.0 (black lines) nm are shown. The following spectra denoted with the number of the residue at which the spin label is attached are spin-normalized ESR spectra (T = 160 K) of singly and doubly spin-labeled mutants of the NpHtrII truncated at position 157 in complex with the NpSRII reconstituted in purple membrane lipids. The ESR spectra (black line) are compared with reference spectra without dipolar broadening (gray line). The strength of dipolar interaction and the resulting line broadening is visible by the decrease in amplitude of the central line with respect to the reference spectra. At the bottom, the DEER time domain data obtained after subtraction of the decay curve due to a 3D homogeneous distribution of spins for spin-labeled position 158 in the F-helix of NpSRII is depicted. To dilute the spin system under analysis, and focus mainly on the intra-dimer interactions, a high lipid/protein ratio has been chosen for this sample. Under this condition it is possible to suppress peaks otherwise present at distances >3.5 nm in the usual lipid/protein ratio used for cw experiments (40-fold excess). The distance distribution is obtained by Tikhonov regularization in the range from 2 to 8 nm with a regularization parameter of 100 (Jeschke 2002). The mean distance of 2.6 nm obtained between the two nitroxides attached at


25

residue S158 in the 2:2 complex agrees with the distance calculated from the crystal structure between the oxygen atoms of the respective serine residues (3 nm). These results once more confirm the presence of a 2:2 complex in membrane. Moreover, they show that DEER experiments can be performed to screen protein/lipid ratio conditions able to induce further aggregation of the complex. (B) Schematic model of the transmembrane region of the proposed 2:2 complex of NpSRII with its transducer NpHtrII (view from the cytoplasm) (Wegener et al. 2001a). The structural model of NpSRII is based on a BR-crystal structure (Essen et al. 1998), validated for side chains 157–160 (Fhelix) and 210–213 (G-helix). Transmembrane helices TM1 and TM2 of NpHtrII are modeled as canonical α-helices. These components are oriented with respect to each other based on dipolar spin–spin interactions and derived distance constraints. Cβ atoms from selected residues are represented as balls. The dotted lines connecting them highlight selected distances derived from ESR data. The continuous line denotes the strong interaction occurring between positions 78 in TM2 and TM2′. (C) Top view of the HAMP domain model up to position 96 (Bordignon et al. 2005). The NpSRII/NpHtrII transmembrane complex obtained from the x-ray data is depicted in light gray ribbon representation. The AS-1 sequence proposed by ESR data is colored in dark grayscale. Cβ atoms from selected residues are represented as balls. The AS-1 segment from G83 to A94 is proposed to be the helical extension of TM2. To account for the high dynamical properties observed, the residues starting from A95 are represented in an extended conformation, although from the ESR data a highly dynamic helical segment cannot be excluded. A structural kink is proposed at residue G83 that leads to the loose coiled-coil motif suggested. (D) Side view of the model.

membrane reconstituted samples suspended in aqueous buffer is retained in the crystal. Figure 5C shows the structural model suggested from analysis of the membrane adjacent region of the transducer in the 2:2 complex in membrane (the side view of the complex is presented in Figure 5D). The study was carried out determining interspin distances between selected residues in the cytoplasmic edge of helix-F (positions 154, 157, 158) and 8 selected residues in the transducer AS-1 sequence beyond position 82 (positions 88–95). Moreover, being the transducer present as a dimer in the complex under investigation, the distances determined between 17 singly spin-labeled transducer mutants gave information on the relative topology of the two transducers in the complex (Bordignon et al. 2005) (examples of spin normalized low-temperature spectra from selected mutants are presented in Figure 5A). In the 88–94 region of the AS-1 sequence, interspin interactions are detectable starting from position 91 (1.9 nm) and become more evident for positions 93 and 94, where the fitting of the simulated spectra to the experimental ones yields distance values of 1.7–1.8 nm. It is worth noting that the clear periodicity of 3.6 residues revealed by the oxygen accessibility from positions 78 to 89 strongly supports the idea that the first part of the AS-1 sequence is the α-helical extension of TM2 (see §3.2, Figure 3). In the case of a straight continuation of the TM2 helix perpendicular to the membrane, positions 93 and 94 are expected to be located within the interface between the neighboring transducers, but should not show stronger interactions than those observed for positions 89 and 90. The fact that closer contacts

26


are observed for positions 93 and 94 indicates a possible inclination of the membrane adjacent part of the transducers toward each other. Despite the highly dynamic nature of region 95–101 (characterized by very high values of the mobility parameter in the 4–5 mT–1 range, not shown) the two transducers interact with each other in the complex showing distances in the 1.8 nm range. The lack of periodicity in the distances is to be expected due to the dynamic properties of this region. To investigate the relative orientation of the AS-1 sequence of NpHtrII with respect to the cognate NpSRII, distance measurements between residues at the cytoplasmic end of helix-F (157F, 158 F) and positions 88–94 of the transducer were performed. Additionally, proximity relations between 154F, located at the cytoplasmic edge of helix-F in the receptor, and positions 91–95 in the transducer have been determined. It has to be pointed out that the system under investigation is a complex of four proteins and a double mutation involving the receptor and the transducer leads to several spin–spin interactions. In particular, dipolar interactions between both receptor–transducer and transducer–transducer positions can lead to the appearance of dipolar broadening in the low-temperature cw ESR spectra. In order to focus only on the dipolar broadening due to the receptor–transducer interaction, the spectra of the doubly labeled variants must be compared to the spectra of the respective singly labeled transducer. The additional decrease in intensity in the spin-normalized spectra indicates the presence of considerable dipolar interaction between the spin-labeled transducer–receptor positions. The closest distances to 157F are observed for positions 88, 89, and 90. In agreement with these findings, the closest distances to 158F are also observed for positions 88 (Figure 5A) and 89, whereas the double mutant NpSRIIS158R1/NpHtrII-S91R1 shows weak interactions. These observations suggest an α-helical arrangement of the transducer residues where the helical turn 88, 89, and 90 is located in the vicinity of residues K157 and S158 in helix-F. In the case of AS-1 being a straight prolongation of helix TM2, close proximity relations are to be expected between the following positions in the transducer (91–95) and 154F, located one helix turn above 157F and 158F. On the contrary, the spectra obtained for these doubly labeled variants do not show any additional broadening with respect to the singly labeled transducers, thus indicating a more complex topology of the transducer dimer. Additionally, the very similar distances observed between position 158F and positions 88–90 in the AS-1 sequence (1.4–1.6 nm) and 80–81 located two helical turns down in TM2 (1.4–1.5 nm) (Wegener et al. 2001a) are also a clear indication that the AS-1 sequence is not the straight prolongation of TM2; otherwise, one would have expected much closer interactions between S158R1 and the 88–90 region than those actually detected. Hence, the AS-1 region following residue S91 is proposed to be in closer contact with the neighboring transducer rather than to the receptor, corroborating the conclusions drawn from the distance values determined from the singly labeled transducer molecules. The AS-1 sequence investigation in terms of interspin interaction allows the proposal of a topological model for this region in the 2:2 complex. Overall, these findings strongly suggest the inclination of the transducers in the membrane adja-


27

cent region to a direction opposite to the receptor, which lead to formation of a loose coiled0coil motif (Figure 5C–D).

4. DETECTION OF CONFORMATIONAL CHANGES

One attractive feature of site-directed spin labeling is the ability to time resolve changes in any of the parameters discussed above. Hence, changes in the protein secondary structure, protein tertiary fold, or domain movements can be followed with up to 100 μs resolution with conventional ESR instrumentation and detection scheme (field modulation). Important examples found in the literature include the detection of rigid-body helix motion in both rhodopsin and bacteriorhodopsin (Farahbakhsh et al. 1993; Farrens et al. 1996; Steinhoff et al. 1994; Thorgeirsson et al. 1997), domain movements in T4 lysozyme (Hubbell et al. 1996), structural reorganization in colicin E1 upon membrane binding (Shin et al. 1993), and conformational changes during signal transfer from sensory rhodopsin NpSRII to the transducer NpHtrII (Wegener et al. 2001a). The strategy for triggering, detecting, and interpreting conformational changes is illustrated with two examples. The first shows detection of a conformational change occurring in rhodopsin upon photoisomerization of the retinal. The second presents the time-resolved detection of changes in mobility and interspin distances in NpSRII and NpHtrII upon light excitation and subsequent isomerization of the retinal, revealing the magnitude and entity of helix motion and signal transfer process. Rhodopsin, shown in the resting state (Figure 6A, dark gray) is the photoreceptor protein of vertebrate retina. Photoisomerization of the bound retinal chromophore triggers a rigid-body helix tilt of a single helix, as demonstrated by a 0.6nm increase in the distance between nitroxide spin labels attached to sites 63 and 241 (Figure 6A, light gray). The interspin distance was measured in the resting and stabilized activated states at cryogenic temperature using DEER (Jeschke 2002). The dipolar evolution functions and the corresponding distance distributions are given for the resting and light-activated states (Figure 6A, right panel). The interspin distance calculated from the crystal structure of the resting state is shown to be in agreement with the ESR determined distance (Altenbach and Hubbel, unpublished). If the protein conformations of interest are of transient nature, the determination of the properties of the respective trapped intermediate states can be accompanied by direct detection of the kinetic difference spectra under physiological conditions. This is illustrated in the case of an NpSRII/NpHtrII complex reconstituted in membrane (Figure 6B). Difference spectra between the photoactivated state and the dark-adapted state can be determined during a single B-field scan by subtraction of

28


Figure 6. (A) Resting (dark gray) and activated (light gray) states of rhodopsin. The two spin labels attached to positions 63 and 241 are shown in stick representation. In the right panel the dipolar evolution functions and the corresponding distances are presented (dark gray: resting state; light gray: activated state). The distance distribution calculated from the crystal structure is also presented as a dotted line, and it is shown to be in agreement with the DEER data. The increase in the distance between positions 63 and 241 by 0.6 nm is clear evidence for a dramatic rearrangement of helix TM6 after excitation. Based on ESR data, a rigid-body helix tilt of TM6 is proposed in the activated state of rhodopsin (Altenbach and Hubbell, unpublished). (B) View onto the cytoplasmic side of the sensory rhodopsin/transducer complex. For sake of clarity, only half of the 2:2 complex is presented, consisting of one NpSRII and one NpHtrII moieties. The two MTS spin labels attached to positions 159F and 78TM2 in helix-F of the receptor and in the TM2 helix of the transducer, respectively, are shown in stick representation. The upper right panel shows the cw ESR spectrum of the spin label attached at position 159F in the complex reconstituted in purple membrane lipids. The spin label at position 159F serves as reporter group for transient structural changes. The lightminus-dark difference spectrum (stick representation) is calculated from the transient ESR traces obtained for each B field position investigated (to allow comparison with the cw spectrum, the difference spectrum is multiplied by 100). Light excitation is achieved via an excimer-pumped dye laser tuned at 500 nm and the ESR transients are recorded for 4 seconds after light excitation, with a pre-trigger of 100 ms. To obtain a good signal-to-noise ratio a lock in time constant of 5 ms is used and an average of 500–1000 ESR transients is performed for each B field value with B field intervals of 0.1 mT. The pre-trigger signal intensity is used to obtain the cw spectrum presented. The height of each stick of the difference spectrum represents the amplitude of the ESR transient, obtained by fitting with a single exponential decay. The difference spectrum thus obtained is found in


29

agreement with previous published difference spectra utilizing the methods of subtraction of signals averaged within two succeeding sampling intervals starting with the light flash of a lamp (Wegener et al. 2000, 2001a). The right bottom panel depicts ESR transients (noisy lines) and the corresponding optical traces (continuous lines) recorded at 500 (ground state), 550 (O state), and 400 nm (M state) of NpSRII/NpHtrII reconstituted in purple membrane (Klare et al. 2004b). The ESR transients are recorded at fixed B field values where a maximum of intensity is observed for NpSRII/NpHtrIIV78R1 (black trace), and for NpSRII-L159R1/NpHtrII (gray trace) after activation with light. The time constant of the lock in is 100 μs, to allow detection of the rise time of the structural changes monitored by ESR. The structural rearrangement clearly occurs synchronously in the receptor and in the transducer within the M1→M2 transition (spectroscopically silent) and it is characterized by a 3 ms time constant. Return to the dark-adapted structure follows recovery of the ground state for position 159, but it is delayed for the transducer. The transient increase of mobility (inferred by the difference spectrum) of the spin label side chain at position 159 suggests an outward displacement of helix-F of the receptor. The ESR signal changes observed for position 78 in the complex uncover a transient distance increase between positions 78 of the two transmembrane helices of the transducer, TM2 and TM2′, as inferred from analysis of the low-temperature spectra of the M-trapped state shown on the left (Wegener et al. 2001a).

the signals averaged within two succeeding sampling intervals starting with the light flash. The lengths of the sampling intervals are set to values longer than the recovery time of the initial state. Following this scheme, the signal recorded during the first sampling interval is an average of the signals of all photo-intermediates including the initial state. The signal determined during the second sampling interval represents the recovered pure initial state. A second more quantitative approach considers the photocycle time constants obtained via optical transient spectroscopy and fit the ESR data utilizing the same set of time constants to allow direct comparison between the optical cycle and the structural changes occurring. The difference spectra reveal the entity and sign of the mobility changes occurring, as for example, in the case of 159F, where a transient increase in the mobility is observed upon activation of NpSRII by a light flash (Figure 6B, upper right panel). The conformational changes occurring in NpSRII upon light excitation observed in the cytoplasmic edge of helix-F are transferred also to the transmembrane part of helix TM2 in the transducer, as exemplified by the ESR transient traces obtained for positions 159F and 78TM2 (Figure 6B, lower right panel). Comparison of the ESR with the optical traces recorded at different wavelength identifies M as the signaling state where the observed conformational changes occur. With the reformation of the ground state, the reactions characteristic for the receptor seems to be decoupled from those of the transducer. The resetting movement of helix-F into the original position seems to precede the recovery of the TM2 position. To further refine the structural changes inferred by the kinetic data, lowtemperature ESR spectra were recorded for several receptor/transducer and transducer/transducer positions in the dark-adapted and M-trapped states revealing distance changes occurring upon light activation (Figure 6B). The data obtained suggest a model of signal transduction where the light-induced F-helix displacement in NpSRII leads to a rotation/screw-like motion of helix TM2 in the transducer. The recently published x-ray structure of the late M state in NpSRII/NpHtrII indeed

30


shows a rotation of TM2 in the signaling state, according to the ESR interpretation (Moukhametzianov et al. 2006). 5. ACKNOWLEDGMENTS

We gratefully acknowledge the support of the Deutsche Forschungsgemeinschaft in the frame of the Schwerpunktprogramm SPP1051 and the Sonderforschungs-bereich SFB 431. We thank W. L. Hubbell for providing us the data shown in Figure 6A. 6. REFERENCES Altenbach C, Flitsch SL, Khorana HG, Hubbell WL. 1989a. Structural studies on transmembrane proteins, 2: spin labeling of bacteriorhodopsin mutants at unique cysteines. Biochemistry 28:7806–7812. Altenbach C, Froncisz W, Hyde JS, Hubbell WL. 1989b. Conformation of spin-labeled melittin at membrane surfaces investigated by pulse saturation recovery and continuous wave power saturation electron-paramagnetic resonance. Biophys J 56:1183–1191. Altenbach C, Greenhalgh DA, Khorana HG, Hubbell WL. 1994. A collision gradient method to determine the immersion depth of nitroxides in lipid bilayers: application to spin-labeled mutants of bacteriorhodopsin. Proc Natl Acad Sci USA 91:1667–1671. Altenbach C, Yang K, Farrens DL, Farahbakhsh ZT, Khorana HG, Hubbell WL. 1996. Structural features and light-dependent changes in the cytoplasmic interhelical E-F loop region of rhodopsin: a site-directed spin-labeling study. Biochemistry 35:12470–12478. Altenbach C, Oh KJ, Trabanino RJ, Hideg K, Hubbell WL. 2001. Estimation of inter-residue distances in spin labeled proteins at physiological temperatures: experimental strategies and practical limitations. Biochemistry 40:15471–15482. Altenbach C, Froncisz W, Hemker R, Mchaourab H, Hubbell WL. 2005. Accessibility of nitroxide side chains: absolute Heisenberg exchange rates from power saturation EPR. Biophys J 89:2103–2112. Anderson DJ, Hanson P, McNulty J, Millhauser G, Monaco V, Formaggio F, Crisma M, Toniolo C. 1999. Solution structures of TOAC-labeled trichogin GA IV peptides from allowed (g approximate to 2) and half-field electron spin resonance. J Am Chem Soc 121:6919–6927. Appleman JA, Chen LL, Stewart V. 2003. Probing conservation of HAMP linker structure and signal transduction mechanism through analysis of hybrid sensor kinases. J Bacteriol 185:4872–4882. Ayers B, Blaschke UK, Camarero JA, Cotton GJ, Holford M, Muir TW. 1999. Introduction of unnatural amino acids into proteins using expressed protein ligation. Biopolymers 51:343–354. Balog M, Kalai T, Jeko J, Berente Z, Steinhoff HJ, Engelhard M, Hideg K. 2003. Synthesis of new conformationally rigid paramagnetic alpha-amino acids. Tetrahedron Lett 44:9213–9217. Balog MR, Kalai TK, Jeko J, Steinhoff HJ, Engelhard M, Hideg K. 2004. Synthesis of new 2,2,5,5-tetramethyl-2,5-dihydro-1H-pyrrol-1-yloxyl radicals and 2-substituted-2,5,5trimethylpyrrolidin-1-yloxyl radicals based alpha-amino acids. Synlett 14:2591–2593.


31

Barbosa SR, Cilli EM, Lamy-Freund MT, Castrucci AML, Nakaie CR. 1999. First synthesis of a fully active spin-labeled peptide hormone. FEBS Lett 446:45–48. Barnes JP, Liang ZC, Mchaourab HS, Freed JH, Hubbell WL. 1999. A multifrequency electron spin resonance study of T4 lysozyme dynamics. Biophys J 76:3298–3306. Becker CFW, Hunter CL, Seidel R, Kent SBH, Goody RS, Engelhard M. 2003. Total chemical synthesis of a functional interacting protein pair: the protooncogene H-Ras and the Ras-binding domain of its effector c-Raf1. Proc Natl Acad Sci USA 100:5075– 5080. Becker CFW, Lausecker K, Balog M, Kalai T, Hideg K, Steinhoff HJ, Engelhard M. 2005. Incorporation of spin-labelled amino acids into proteins. Magn Reson Chem 43:S34– S39. Berliner LJ, ed. 1976. Spin labeling: theory and applications. New York: Academic Press. Berliner LJ, ed. 1979. Spin labeling ii: theory and applications. New York: Academic Press. Berliner LJ, Reuben J, eds. 1989. Spin labeling theory and applications, Vol. 8: Biological magnetic resonance. New York: Plenum Press. Berliner LJ, Grunwald J, Hankovszky HO, Hideg K. 1982. A novel reversible thiol-specific spin label: papain active-site labeling and inhibition. Anal Biochem 119:450–455. Berliner LJ, Eaton SS, Eaton GR, eds. 2000. Distance measurements in biological systems by EPR: biological magnetic resonance. New York: Kluwer Academic/Plenum. Bolin KA, Anderson DJ, Trulson JA, Thompson DA, Wilken J, Kent SBH, Gantz I, Millhauser GL. 1999. NMR structure of a minimized human agouti related protein prepared by total chemical synthesis. FEBS Lett 451:125–131. Borbat PP, Freed JH. 1999. Multiple-quantum ESR and distance measurements. Chem Phys Lett 313:145–154. Borbat PP, Costa-Filho AJ, Earle KA, Moscicki JK, Freed JH. 2001. Electron spin resonance in studies of membranes and proteins. Science 291:266–269. Bordignon E, Klare JP, Doebber M, Wegener AA, Martell S, Engelhard M, Steinhoff HJ. 2005. Structural analysis of a HAMP domain: the linker region of the phototransducer in complex with sensory rhodopsin II. J Biol Chem 280:38767–38775. Burghaus O, Rohrer M, Gotzinger T, Plato M, Möbius K. 1992. A novel high-field highfrequency EPR and ENDOR spectrometer operating at 3 mm wavelength. Meas Sci Technol 3:765–774. Cai KW, Langen R, Hubbell WL, Khorana HG. 1997. Structure and function in rhodopsin: topology of the C-terminal polypeptide chain in relation to the cytoplasmic loops. Proc Natl Acad Sci USA 94:14267–14272. Chin JW, Cropp TA, Anderson JC, Mukherji M, Zhang ZW, Schultz PG. 2003. An expanded eukaryotic genetic code. Science 301:964–967. Closs GL, Forbes MDE, Piotrowiak P. 1992. Spin and reaction dynamics in flexible polymethylene biradicals as studied by EPR, NMR, and optical spectroscopy and magneticfield effects: measurements and mechanisms of scalar electron–spin spin coupling. J Am Chem Soc 114:3285–3294. Cornish VW, Benson DR, Altenbach CA, Hideg K, Hubbell WL, Schultz PG. 1994. Sitespecific incorporation of biophysical probes into proteins. Proc Natl Acad Sci USA 91:2910–2914. Cuello LG, Cortes DM, Perozo E. 2004. Molecular architecture of the KvAP voltagedependent K+ channel in a lipid bilayer. Science 306:491–495. Dawson PE, Kent SBH. 2000. Synthesis of native proteins by chemical ligation. Annu Rev Biochem 69:923–960.

32


Elsaber C, Monien B, Haehnel W, Bittl R. 2005. Orientation of spin labels in de novo peptides. Magn Reson Chem 43:S26–S33. Essen LO, Siegert R, Lehmann WD, Oesterhelt D. 1998. Lipid patches in membrane protein oligomers: crystal structure of the bacteriorhodopsin-lipid complex. Proc Natl Acad Sci USA 95:11673–11678. Farahbakhsh ZT, Altenbach C, Hubbell WL. 1992. Spin labeled cysteines as sensors for protein lipid interaction and conformation in rhodopsin. Photochem Photobiol 56:1019–1033. Farahbakhsh ZT, Hideg K, Hubbell WL. 1993. Photoactivated conformational changes in rhodopsin: a time-resolved spin-label study. Science 262:1416–1419. Farrens DL, Altenbach C, Yang K, Hubbell WL, Khorana HG. 1996. Requirement of rigidbody motion of transmembrane helices for light activation of rhodopsin. Science 274:768–770. Feix JB, Klug CS. 1998. Site directed spin labeling of membrane proteins and peptidemembrane interactions. In Spin labeling: the next millennium, Vol. 14, pp. 251–281. Ed LJ Berliner. New York: Plenum. Feix JB, Popp CA, Venkataramu SD, Beth AH, Park JH, Hyde JS. 1984. An electron– electron double-resonance study of interactions between [N-14]stearic and [N15]stearic acid spin-label pairs: lateral diffusion and vertical fluctuations in dimyristoylphosphatidylcholine. Biochemistry 23:2293–2299. Fiori WR, Millhauser GL. 1995. Exploring the peptide 3(10)-helix-reversible-arrow-alphahelix equilibrium with double-label electron-spin-resonance. Biopolymers 37:243–250. Freed JH. 1976. Theory of slow tumbing ESR spectra for nitroxides. In Spin labeling: theory and applications, pp. 53–132. Ed LJ Berliner. New York: Academic Press. Gordeliy VI, Labahn J, Moukhametzianov R, Efremov R, Granzin J, Schlesinger R, Büldt G, Savopol T, Scheidig AJ, Klare JP, Engelhard M. 2002. Molecular basis of transmembrane signalling by sensory rhodopsin II-transducer complex. Nature 419:484–487. Griffith OH, Dehlinger PJ, Van SP. 1974. Shape of hydrophobic barrier of phospholipid bilayers (evidence for water penetration in biological-membranes). J Membr Biol 15:159–192. Gross A, Columbus L, Hideg K, Altenbach C, Hubbell WL. 1999. Structure of the KcsA potassium channel from Streptomyces lividans: a site-directed spin labeling study of the second transmembrane segment. Biochemistry 38:10324–10335. Hahn ME, Muir TW. 2005. Manipulating proteins with chemistry: a cross-section of chemical biology. Trends Biochem Sci 30:26–34. Hanson P, Martinez G, Millhauser G, Formaggio F, Crisma M, Toniolo C, Vita C. 1996a. Distinguishing helix conformations in alanine-rich peptides using the unnatural amino acid TOAC and electron spin resonance. J Am Chem Soc 118:271–272. Hanson P, Millhauser G, Formaggio F, Crisma M, Toniolo C. 1996b. ESR characterization of hexameric, helical peptides using double TOAC spin labeling. J Am Chem Soc 118:7618–7625. Hubbell WL, Mchaourab HS, Altenbach C, Lietzow MA. 1996. Watching proteins move using site-directed spin labeling. Structure 4:779–783. Hubbell WL, Gross A, Langen R, Lietzow MA. 1998. Recent advances in site-directed spin labeling of proteins. Curr Opin Struct Biol 8:649–656. Hubbell WL, Cafiso DS, Altenbach C. 2000. Identifying conformational changes with sitedirected spin labeling. Nat Struct Biol 7:735–739.


33

Huber M, Törring JT. 1995. High-field EPR on the primary electron-donor cation-radical in single-crystals of heterodimer mutant reaction centers of photosynthetic bacteria: first characterization of the g-tensor. Chem Phys 194:379–385. Hustedt EJ, Beth AH. 1999. Nitroxide spin–spin interactions: applications to protein structure and dynamics. Ann Rev Biophys Biomol Struct 28:129–153. Hustedt EJ, Smirnov AI, Laub CF, Cobb CE, Beth AH. 1997. Molecular distances from dipolar coupled spin-labels: the global analysis of multifrequency continuous wave electron paramagnetic resonance data. Biophys J 72:1861–1877. Isas JM, Langen R, Haigler HT, Hubbell WL. 2002. Structure and dynamics of a helical hairpin and loop region in annexin 12: a site-directed spin labeling study. Biochemistry 41:1464–1473. Jeschke G. 2002. Distance measurements in the nanometer range by pulse EPR. Chem Phys Chem 3:927–932. Kim KK, Yokota H, Kim SH. 1999. Four-helical-bundle structure of the cytoplasmic domain of a serine chemotaxis receptor. Nature 400:787–792. Klare JP, Bordignon E, Engelhard M, Steinhoff HJ. 2004a. Sensory rhodopsin II and bacteriorhodopsin: light activated helix F movement. Photochem Photobiol Sci 3:543–547. Klare JP, Gordeliy VI, Labahn J, Büldt G, Steinhoff HJ, Engelhard M. 2004b. The archaeal sensory rhodopsin II/transducer complex: a model for transmembrane signal transfer. FEBS Lett 564:219–224. Kochendoerfer GG, Chen SY, Mao F, Cressman S, Traviglia S, Shao HY, Hunter CL, Low DW, Cagle EN, Carnevali M, Gueriguian V, Keogh PJ, Porter H, Stratton SM, Wiedeke MC, Wilken J, Tang J, Levy JJ, Miranda LP, Crnogorac MM, Kalbag S, Botti P, Schindler-Horvat J, Savatski L, Adamson JW, Kung A, Kent SBH, Bradburne JA. 2003. Design and chemical synthesis of a homogeneous polymer-modified erythropoiesis protein. Science 299:884–887. Kochendoerfer GG, Jones DH, Lee S, Oblatt-Montal M, Opella SJ, Montal M. 2004. Functional characterization and NMR Spectroscopy on full-length Vpu from HIV-1 prepared by total chemical synthesis. J Am Chem Soc 126:2439–2446. Koteiche HA, Mchaourab HS. 1999. Folding pattern of the alpha-crystallin domain in alpha A-crystallin determined by site-directed spin labeling. J Mol Biol 294:561–577. Koteiche HA, Berengian AR, Mchaourab HS. 1998. Identification of protein folding patterns using site-directed spin labeling: structural characterization of a beta-sheet and putative substrate binding regions in the conserved domain of alpha A-crystallin. Biochemistry 37:12681–12688. LaConte LEW, Voelz V, Nelson W, Enz M, Thomas DD. 2002. Molecular dynamics simulation of site-directed spin labeling: experimental validation in muscle fibers. Biophys J 83:1854–1866. Likhtenshtein GI. 1976. Spin labeling methods in molecular biology. New York: John Wiley & Sons. Luecke H, Richter HT, Lanyi JK. 1998. Proton transfer pathways in bacteriorhodopsin at 2.3 Å resolution. Science 280:1934–1937. Luecke H, Schobert B, Richter HT, Cartailler JP, Lanyi JK. 1999. Structure of bacteriorhodopsin at 1.55 Å resolution. J Mol Biol 291:899–911. Marchetto R, Schreier S, Nakaie CR. 1993. A novel spin-labeled amino-acid derivative for use in peptide-synthesis: (9-Fluorenylmethyloxycarbonyl)-2,2,6,6tetramethylpiperidine-N-oxyl-4-amino-4-carboxylic acid. J Am Chem Soc 115:11042– 11043.

34


Marsh D, Dzikovski BG, Livshits VA. 2006. Oxygen profile in membranes. Biophys J 90:L49–L51. Mchaourab HS, Hyde JS. 1993. Dependence of the multiple-quantum EPR signal on the spin-lattice relaxation-time: effect of oxygen in spin-labeled membranes. J Magn Reson Ser B 101:178–184. Mchaourab HS, Perozo E. 2000. Determination of protein folds and conformational dynamics using spin labeling EPR spectroscopy. In Distance measurements in biological systems by EPR, pp. 155–218. Ed LJ Berliner, SS Eaton, GR Eaton. New York: Kluwer. Mchaourab HS, Lietzow MA, Hideg K, Hubbell WL. 1996. Motion of spin-labeled side chains in T4 lysozyme, correlation with protein structure and dynamics. Biochemistry 35:7692–7704. Mendel D, Cornish VW, Schultz PG. 1995. Site-directed mutagenesis with an expanded genetic-code. Ann Rev Biophys Biomol Struct 24:435–462. Merrifield RB. 1963. Solid phase peptide synthesis, 1: synthesis of a tetrapeptide. J Am Chem Soc 85:2149–2154. Möbius K, Savitsky A, Wegener C, Plato M, Fuchs M, Schnegg A, Dubinskii AA, Grishin YA, Grigor'ev IA, Kühn M, Duché D, Zimmermann H, Steinhoff HJ. 2005. Combining high-field EPR with site-directed spin labeling reveals unique information on proteins in action. Magn Reson Chem 43:S4–S19. Monaco V, Formaggio F, Crisma M, Toniolo C, Hanson P, Millhauser G, George C, Deschamps JR, Flippen-Anderson JL. 1999. Determining the occurrence of a 3(10)helix and an alpha-helix in two different segments of a lipopeptaibol antibiotic using TOAC, a nitroxide spin-labeled C-alpha-tetrasubstituted alpha-amino acid. Bioorg Med Chem 7:119–131. Moukhametzianov R, Klare JP, Efremov R, Baeken C, Göppner A, Labahn J, Engelhard M, Büldt G, Gordeliy VI. 2006. Development of the signal in sensory rhodopsin and its transfer to the cognate transducer. Nature 440:115–119. Nielsen RD, Canaan S, Gladden JA, Gelb MH, Mailer C, Robinson BH. 2004. Comparing continuous wave progressive saturation EPR and time domain saturation recovery EPR over the entire motional range of nitroxide spin labels. J Magn Reson 169:129–163. O'Donnell MJ, Bennett WD, Wu SD. 1989. The stereoselective synthesis of alpha-aminoacids by phase-transfer catalysis. J Am Chem Soc 111:2353–2355. Pannier M, Veit S, Godt A, Jeschke G, Spiess HW. 2000. Dead-time free measurement of dipole–dipole interactions between electron spins. J Magn Reson 142:331–340. Pebay-Peyroula E, Rummel G, Rosenbusch JP, Landau EM. 1997. X-ray structure of bacteriorhodopsin at 2.5 angstroms from microcrystals grown in lipidic cubic phases. Science 277:1676–1681. Percival PW, Hyde JS. 1975. Pulsed EPR spectrometer, 2. Rev Sci Instrum 46:1522–1529. Perozo E, Cortes DM, Cuello LG. 1998. Three-dimensional architecture and gating mechanism of a K+ channel studied by EPR spectroscopy. Nat Struct Biol 5:459–469. Pfeiffer M, Rink T, Gerwert K, Oesterhelt D, Steinhoff HJ. 1999. Site-directed spin-labeling reveals the orientation of the amino acid side-chains in the E-F loop of bacteriorhodopsin. J Mol Biol 287:163–171. Plato M, Steinhoff HJ, Wegener C, Törring JT, Savitsky A, Möbius K. 2002. Molecular orbital study of polarity and hydrogen bonding effects on the g and hyperfine tensors of site directed NO spin labelled bacteriorhodopsin. Mol Phys 100:3711–3721. Poole CP. 1983. Electron spin resonance. New York: Wiley.


35

Prisner TF, Vanderest A, Bittl R, Lubitz W, Stehlik D, Möbius K. 1995. Time-resolved Wband (95 Ghz) EPR spectroscopy of Zn-substituted reaction centers of rhodobactersphaeroides R-26. Chem Phys 194:361–370. Pyka J, Ilnicki J, Altenbach C, Hubbell WL, Froncisz W. 2005. Accessibility and dynamics of nitroxide side chains in T4 lysozyme measured by saturation recovery EPR. Biophys J 89:2059–2068. Qin PZ, Hideg K, Feigon J, Hubbell WL. 2003. Monitoring RNA base structure and dynamics using site-directed spin labeling. Biochemistry 42:6772–6783. Rabenstein MD, Shin YK. 1995. Determination of the distance between 2 spin labels attached to a macromolecule. Proc Natl Acad Sci USA 92:8239–8243. Radzwill N, Gerwert K, Steinhoff HJ. 2001. Time-resolved detection of transient movement of helices F and G in doubly spin-labeled bacteriorhodopsin. Biophys J 80:2856–2866. Rassat A, Rey P. 1967. Nitroxides, 23: preparation of amino-acid free radicals and their complex salts. Bull Soc Chim Fr 3:815–818. Sale K, Song LK, Liu YS, Perozo E, Fajer P. 2005. Explicit treatment of spin labels in modeling of distance constraints from dipolar EPR and DEER. J Am Chem Soc 127:9334– 9335. Shafer AM, Kalai T, Liu SQB, Hideg K, Voss JC. 2004. Site-specific insertion of spinlabeled L-amino acids in Xenopus oocytes. Biochemistry 43:8470–8482. Shin YK, Hubbell WL. 1992. Determination of electrostatic potentials at biological interfaces using electron electron double-resonance. Biophys J 61:1443–1453. Shin YK, Levinthal C, Levinthal F, Hubbell WL. 1993. Colicin-E1 binding to membranes: time-resolved studies of spin-labeled mutants. Science 259:960–963. Steinhoff HJ. 2002. Methods for study of protein dynamics and protein–protein interaction in protein ubiquitination by electron paramagnetic resonance spectroscopy. Front Biosci 7:C97–C110. Steinhoff HJ, Hubbell WL. 1996. Calculation of electron paramagnetic resonance spectra from Brownian dynamics trajectories: application to nitroxide side chains in proteins. Biophys J 71:2201–2212. Steinhoff HJ, Karim C. 1993. Protein dynamics and EPR-spectroscopy: comparison of molecular dynamic simulations with experiments. Ber Bunsenges Phys Chem 97:163–171. Steinhoff HJ, Lieutenant K, Schlitter J. 1989. Residual motion of hemoglobin-bound spin labels as a probe for protein dynamics. Z Naturforschung C: J Biosci 44:280–288. Steinhoff HJ, Dombrowsky O, Karim C, Schneiderhahn C. 1991. Two-dimensional diffusion of small molecules on protein surfaces: an EPR study of the restricted translational diffusion of protein-bound spin labels. Eur Biophys J Biophy Lett 20:293–303. Steinhoff HJ, Mollaaghababa R, Altenbach C, Hideg K, Krebs M, Khorana HG, Hubbell WL. 1994. Time-resolved detection of structural changes during the photocycle of spinlabeled bacteriorhodopsin. Science 266:105–107. Steinhoff HJ, Radzwill N, Thevis W, Lenz V, Brandenburg D, Antson A, Dodson G, Wollmer A. 1997. Determination of interspin distances between spin labels attached to insulin: Comparison of electron paramagnetic resonance data with the x-ray structure. Biophys J 73:3287–3298. Steinhoff HJ, Pfeiffer M, Rink T, Burlon O, Kurz M, Riesle J, Heuberger E, Gerwert K, Oesterhelt D. 1999. Azide reduces the hydrophobic barrier of the bacteriorhodopsin proton channel. Biophys J 76:2702–2710. Steinhoff HJ, Müller M, Beier C, Pfeiffer M. 2000a. Molecular dynamics simulation and EPR spectroscopy of nitroxide side chains in bacteriorhodopsin. J Mol Liq 84:17–27.

36


Steinhoff HJ, Savitsky A, Wegener C, Pfeiffer M, Plato M, Möbius K. 2000b. High-field EPR studies of the structure and conformational changes of site-directed spin labeled bacteriorhodopsin. Biochim Biophys Acta Bioenerg 1457:253–262. Stone AJ. 1963. Gauge invariance of g tensor. Proc Roy Soc London A 271:424–434. Thorgeirsson TE, Xiao WZ, Brown LS, Needleman R, Lanyi JK, Shin YK. 1997. Transient channel-opening in bacteriorhodopsin: an EPR study. J Mol Biol 273:951–957. Voet D, Voet JG. 2004. Biochemistry, 3rd ed. New York: John Wiley & Sons. Voss J, Hubbell WL, Kaback HR. 1998. Helix packing in the lactose permease determined by metal–nitroxide interaction. Biochemistry 37:211–216. Wegener AA, Chizhov I, Engelhard M, Steinhoff HJ. 2000. Time-resolved detection of transient movement of helix F in spin-labelled pharaonis sensory rhodopsin II. J Mol Biol 301:881–891. Wegener AA, Klare JP, Engelhard M, Steinhoff HJ. 2001a. Structural insights into the early steps of receptor-transducer signal transfer in archaeal phototaxis. EMBO J 20:5312– 5319. Wegener C, Savitsky A, Pfeiffer M, Möbius K, Steinhoff HJ. 2001b. High-field EPRdetected shifts of magnetic tensor components of spin label side chains reveal protein conformational changes: the proton entrance channel of bacteriorhodopsin. Appl Magn Reson 21:441–452. Yin JJ, Pasenkiewiczgierula M, Hyde JS. 1987. Lateral diffusion of lipids in membranes by pulse saturation recovery electron-spin-resonance. Proc Natl Acad Sci USA 84:964–968.